Summary

Currently Astrolabe uses salt + timestamp to produce the project name. This can cause multiple tests running at the same time to get the same exact project name, which would subsequently cause Astrolabe to attempt to create the same project twice and fail.

A random value should be added to the generated project name for each test run to avoid this issue.

This manifests as the following error:

[2022/06/08 05:12:28.972] atlasclient.exceptions.AtlasApiError: 404: Not Found. Error Code: 'NOT_ATLAS_GROUP': Group 62a02fbc189fa56883c86716 is not an Atlas group; use the Cloud Manager Public API at /api/public/v1.0 to access it. (GET https://cloud-qa.mongodb.com/api/atlas/v1.0/groups/byName/drivers-atlas-testing-1654665148)

https://evergreen.mongodb.com/task_log_raw/drivers_atlas_testing_tests_dotnet_windows__driver~dotnet_master_platform~windows_64_runtime~dotnet_async_netcoreapp2.1_failure_write_62a02cd157e85a577146cbd3_22_06_08_05_00_01/0?type=T&text=true


Motivation

Who is the affected end user?

Driver developers.

How does this affect the end user?

Users see test failures when there is no genuine failure to solve.

How likely is it that this problem or use case will occur?

The problem occurs rather frequently because many tests are scheduled to start at the same exact time (20+ across all languages).

If the problem does occur, what are the consequences and how severe are they?

The problem causes spurious test failures.

Is this issue urgent?

No

Is this ticket required by a downstream team?

No

Is this ticket only for tests?

No